On Ceva’s and Menelaus’s Theorems for a Möbius triangle

Ivan Livinsky · 2026

Read On Ceva’s and Menelaus’s Theorems for a Möbius triangle by Ivan Livinsky (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-29905.txt.

Introduction

Ivan Livinsky extends the classical theorems of Ceva and Menelaus to curvilinear triangles bounded by circular arcs. He introduces trilinear coordinates tied to such triangles and builds several geometric constructions upon them. For any proper Möbius triangle he defines the incenter, the excenters, and the orthocenter, the classic triangle centers carried into the curved setting. The work thus moves two pillars of triangle geometry beyond straight edges, giving arc-bounded triangles their own center theory.

Catalog record

CREATORIvan Livinsky
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE114839
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.29905

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

On Ceva’s and Menelaus’s Theorems for a Möbius triangle Ivan Livinsky Abstract We generalize the classical Ceva’s and Menelaus’s theorems to curvilinear triangles bounded by circular arcs (Figure 1 ). We introduce trilinear coordinates associated with such triangles and develop several geometric constructions. In particular, for any proper Möbius triangle we define the incenter, excenters, and orthocenter. An interactive demo containing most of our results is available at https://github.com/livinsky-research/Moebius Figure 1: Ceva’s Theorem for a Möbius triangle gives necessary and sufficient conditions for three cevians to belong to one pencil. 1 The Möbius plane The main purpose of this paper is to show that Möbius geometry is no less rich than classical geometries such as affine or hyperbolic. As we shall see, the results obtained here naturally yield corresponding statements for spherical, hyperbolic, and Euclidean geometries, illustrating the unifying role of Möbius geometry. We begin by recalling the basic properties of the Möbius plane. The Möbius plane \mathbfit{M} is an incidence structure of points and cycles . In its classical model we identify \mathbfit{M} with the Riemann sphere \mathbb{C}\cup\{\infty\} (or, equivalently, with the complex projective line \mathbb{C}P^{1} ). In this representation the cycles are all circles and all straight lines with \infty attached. For any three points there exists a unique cycle containing them. Straight lines are exactly the cycles containing \infty . Any cycle \mathbfit{a} can be defined by an equation of the form kz\overline{z}+lz+m\overline{z}+n=0, where the matrix M=\left(\begin{smallmatrix}k&l\\ m&n\end{smallmatrix}\right)\in\mathrm{Herm}(2) is Hermitian, non-degenerate, and indefinite [ 2 ] . In particular, k and n are real, l=\overline{m} , and \det M<0 . Two matrices define the same cycle whenever they differ by a real scalar factor. A cycle is a straight line iff k=0 . When k\neq 0 the cycle is a circle with center O=-\frac{m}{k} and radius r satisfying r^{2}=\frac{|m|^{2}}{k^{2}}-\frac{n}{k}=-\frac{1}{k^{2}}\det M . The Möbius plane has the isometry group \Iso\mathbfit{M} that maps cycles to cycles and preserves incidence. Two sets X,Y\subseteq\mathbfit{M} are congruent if there exists an isometry \sigma\in\Iso\mathbfit{M} such that \sigma(X)=Y . The group \Iso\mathbfit{M} is generated by all inversions \sigma{a} about all cycles \mathbfit{a} . In the case of line the corresponding inversion is the reflection across the line. The isometry group has two connectedness components. The component of the identity is the orientation-preserving subgroup \Iso^{+}\mathbfit{M} . It is a subgroup of index 2 and consists of all Möbius transformations z\mapsto\frac{az+b}{cz+d} and is naturally isomorphic to the group \mathrm{PSL}(2;\mathbb{C}) . The remaining isometries are maps of the form z\mapsto\frac{a\overline{z}+b}{c\overline{z}+d}. In particular, all rotations, translations, homotheties, and similarities are Möbius isometries. The only isometries that preserve \infty are of the form z\mapsto az+b or z\mapsto a\overline{z}+b , for some a,b\in\mathbb{C} , a\neq 0 . These are exactly the Euclidean similarities. The full group \Iso\mathbfit{M} is generated by \Iso^{+}\mathbfit{M} and any inversion \sigma{a} , or by all similarities and any inversion \sigma{a} provided that \mathbfit{a} is a circle. All the isometry transformations are conformal. Therefore, angles are well-defined in Möbius geometry and agree with the angles in the Euclidean representation. On the other hand, there is no distance, no area, and no in-between relation for triples of points; moreover, every cycle has no center. The standard constant-curvature geometries (spherical, Euclidean, and hyperbolic) are naturally embedded into Möbius geometry, and every theorem in Möbius geometry becomes some result in these geometries involving lines and circles. • for spherical geometry we identify \mathbfit{M} with the sphere S^{2} via stereographic projection. The cycles are then all straight lines and all circles. All are obtained as planar sections of the sphere. The planes corresponding to the straight lines are the ones passing through the center of S^{2} . After projecting onto \mathbb{C} the straight lines become the lines passing through 0 and the circles relative to which the power of 0 is -1 . In the matrix language these cycles correspond to the matrices M with an additional property k+n=\trr M=0 . • for Euclidean geometry we consider \mathbb{C}=\mathbfit{M}\setminus\{\infty\} and once again the cycles are all straight lines and all circles. The straight lines are identified by the condition k=0 . • for hyperbolic geometry we consider the upper half-plane \mathcal{H}\subset\mathbb{C} as the Poincaré model. The straight lines are the half-circles orthogonal to \mathbb{R} and vertical rays. These are exactly the cycles defined by real matrices M . The remaining cycles that intersect \mathcal{H} cover all hyperbolic circles, horocycles, and equidistant curves depending on the number of intersection points with \overline{\mathbb{R}}=\mathbb{R}\cup\{\infty\} (this is also the real projective line \mathbb{R}P^{1} ). In all our proofs we are not going beyond complex numbers, trigonometry, basic linear algebra, and quadratic equations. 2 Points and cycles The following property of Möbius transformations is well-known from Complex Analysis. Proposition 2.1 . For any two triples A,B,C and A^{\prime},B^{\prime},C^{\prime} of distinct points on \mathbfit{M} there exists a unique orientation-preserving isometry \sigma\in\Iso^{+}\mathbfit{M} such that \sigma(A)=A^{\prime} , \sigma(B)=B^{\prime} , \sigma(C)=C^{\prime} . It means that any three distinct points on the Möbius plane are in general position. Hence, not only there are no two-point invariants such as distance, but no three-point invariants either. Only for four points we have some special cases and nontrivial invariants. For any two points there are infinitely many cycles passing through them. Two points lying on a cycle split it into two cycle segments. Let A,B,C,D be four distinct points lying on some cycle. We say that A,B split C,D if C and D belong to different segments formed by A and B . For any A,B,C,D we have either A,B split C,D , or A,C split B,D , or A,D split B,C (Figure 2 ). A C D B A,B split C,D A B D C A,C split B,D A B C D A,D split B,C Figure 2: Relative position of four points on a cycle. Furthermore, the cross-ratio (A,B;C,D)=\frac{C-A}{C-B}\frac{D-B}{D-A} is real and is invariant under Möbius isometries. However, we will not need it in this paper. Now consider the relative position of cycles on the Möbius plane. Every cycle splits the plane into two half-planes . Two cycles \mathbfit{a},\mathbfit{b} are either disjoint, intersect in two points, or intersect in one point. In the last case we say that \mathbfit{a} and \mathbfit{b} are tangent . For example, two parallel lines are tangent in \infty . For three pairwise disjoint cycles \mathbfit{a} , \mathbfit{b} , \mathbfit{c} we naturally have the in-between relation. We say that \mathbfit{a} is between \mathbfit{b} and \mathbfit{c} , or \mathbfit{a} splits \mathbfit{b} and \mathbfit{c} if \mathbfit{b} and \mathbfit{c} lie in different half-planes bounded by \mathbfit{a} . Clearly, for any three pairwise disjoint cycles either one of them or none is lying between two others (Figure 3 ). Figure 3: Relative position of three pairwise disjoint cycles. Two disjoint cycles split the plane into two half-planes and an annulus . For example, a ring between two concentric circles is an annulus. Two tangent cycles split the plane into two half-planes and a new region – a monogon – a domain bounded by two cycles and one point – its vertex. For example, a strip between two parallel lines is a monogon with \infty being its vertex. Proposition 2.2 . Any two monogons are congruent. Proof. Let \mathbfit{a},\mathbfit{b} be two tangent cycles. Without loss

Read the complete text